A general algorithm for peak-tracking in multi-dimensional NMR experiments.

We present an algorithmic method allowing automatic tracking of NMR peaks in a series of spectra. It consists in a two phase analysis. The first phase is a local modeling of the peak displacement between two consecutive experiments using distance matrices. Then, from the coefficients of these matrices, a value graph containing the a priori set of possible paths used by these peaks is generated. On this set, the minimization under constraint of the target function by a heuristic approach provides a solution to the peak-tracking problem. This approach has been named GAPT, standing for General Algorithm for NMR Peak Tracking. It has been validated in numerous simulations resembling those encountered in NMR spectroscopy. We show the robustness and limits of the method for situations with many peak-picking errors, and presenting a high local density of peaks. It is then applied to the case of a temperature study of the NMR spectrum of the Lipid Transfer Protein (LTP).

[Structure-mobility relationships in thin-layer chromatography. Applications to some progestogens (author's transl)]. / Relations structure-mobilité en chromatographie sur couches minces. Applications à quelques progestatifs

A study has been made on the quantitative relationship between the chromatographic mobility and the structure of progestogenic steroids. The coefficient K, corresponding to the fundamental skeleton, and the parameters of different substituents were calculated from the RM values. The results obtained with eighteen steroids tested in two solvent systems of low dielectric constant and low polarity have been improved satisfactorily.